This invention is directed to the self-assembly of three-dimensional structures, and more particularly, to the self-assembly of three-dimensional structures for integrated circuits utilizing shaped magnetic fields.
Over the past few decades, advances in the field of electronic and photonic devices for signal processing wireless communication, computing and the like have become more complex as they become more highly integrated. Most systems are an amalgam of heavily integrated microsystems. These systems are enabled by careful integration of subsystems of various physical size and function into compact modular devices. By way of example, the cellular phone tightly integrates CMOS electronics, RF components, photonics, micro-electro-mechanical systems (MEMS), passive discrete components, power supplies and circuit boards to enable wireless communication, data processing/storage and signal processing functionalities. The trend is for smaller, more compact, more integrated and more complex devices.
It is currently known in the art to create these devices by separately fabricating each of the components, through mutually incompatible processes, and then assembling into the sub-assemblies. For example, the semiconductor devices are batch fabricated using wafers having hundreds or thousands of individual components.
Reference is made to FIG. 1A in which a schematic of a known assembly process is shown. Semiconductor devices are batch fabricated using wafers 10. After microfabrication, the wafers are diced and possibly packaged into subcomponents 12. In parallel with the semiconductor processing, passive electrical components such as capacitors, inductors, and resistors, as well as circuit boards 14 are fabricated. After individual fabrication, all of the system subcomponents are assembled serially onto the circuit boards utilizing human manipulation, or more commonly robotic “pick and place” systems to form the completed electrical circuit board 15. This method has been satisfactory; however, it suffers from the disadvantage that the large number of components that can be batch manufactured overwhelm the throughput capabilities for the serial back-end packaging and assembly. Because of the serial nature, the throughput is limited by the number and speed of the robotic manipulators. Secondly, the shrinking physical size of the microelectronics, down to the micro- or nanoscale in many applications, requires precise manipulation. As components become smaller, the absolute alignment positioning tolerances scale equivalently; sometimes beyond the capabilities of robotic manipulation.
Lastly, for sub-millimeter parts, the adhesion forces between the part and the manipulator are significant compared to gravity, resulting in a sticking problem.
Self-assembly has been utilized to a small extent during the packaging and preparation process when forming components onto substrates. Self-assembly is the autonomous organization of components into patterns or structures without human intervention. It is the assembly of parts onto a fixed substrate or the assembly of homogeneous or heterogeneous mixtures of parts one to another. The process is inherently stochastic, relying on the random distribution, mixing and physical interactions between the parts.
The self-assembly processes are governed by two fundamental forces, one being played against the other. Namely, the two forces are the mixing forces causing the large-scale mechanical movement between the parts to be assembled and short-range bonding forces causing the parts to assemble to one another when in close proximity. To maintain a bond, the short range bonding force must be greater than the mixing force to provide a stable connection. The short range bonding force must also be sufficient to overcome other external forces that may act to separate the parts, such as gravity, surface tension, buoyancy, electrostatic forces or the like. Alignment of the parts in self-assembly is typically dependent on minimization of the total free energy.
It is also known in the art to effect mixing by fluid flow (typically for wet assembly) or by vibration energy (typically for dry assembly). Short-range bonding forces have taken the forms of gravity, electrostatic forces, magnetic forces, and capillary forces (as seen in FIG. 2). It should be noted that the forces need merely be sufficient to temporarily hold two pieces of interest together even in the presence of the mixing forces. Once the parts have been assembled, permanent, mechanical and electrical connections can be made through the curing of polymer-based adhesives or reflowing of solder bumps.
More particularly, it is known to use electrostatic forces to more efficiently direct and hold one part to another. In one prior art embodiment, pattern electrodes have been formed on the surface of at least one of the parts to create electric field traps to capture and surface mount components and LEDs on silicon substrates. This method has been satisfactory; however, it requires the formation of an electrical circuit in a desired pattern on the substrate. This adds complexity to the structure of a substrate, is often substrate-limited, i.e., is too difficult to use between two free-floating bodies, and requires the input of energy.
One way of overcoming this shortcoming is to create direct forces between the component and the substrate such as magnetic forces. It is known in the art to use magnetic forces to self-assemble 50 μm nickel disks coated with immobilized biomaterials into an array of nickel disk pattern on the substrate. However, in the prior art, the magnetic forces were the result of the entire structure being a magnet where the structures were attracted to each other, but there is no control over selectivity, or interaction. Furthermore, the magnetic approach was also limited to substrate bonding, not to free floating bodies.
To overcome the shortcomings of these self-assembly structures and methods, capillary forces have also been used to drive self-assembly. At small scales, the capillary forces become dominant. The hydrophilicity of various regions of a surface is controlled to pattern liquid films on a substrate. When a hydrophilic contact pad on one side of a part comes in contact with a liquid droplet, the pad spontaneously wets, and capillary forces draw the part into alignment, thereby minimizing the fluidic interfacial surface energy. This technique has come into vogue to assemble small parts onto planar surfaces with submicron precision for micromirror arrays, inductors and micropumps by way of example.
These techniques and methods for manufacture have been satisfactory; however, they do not match the full functionality offered by robotic or human part manipulation such as orientational uniqueness, bonding selectivity, or inter-part bonding. As a result, they limit the basic logical design rules available for a designer in the self-assembly process.
To provide orientational uniqueness, the method must restrict part bonding to a unique physical orientation between the two bodies to be bonded. As a result, process yield is improved by minimizing the number of misaligned, misfit or incorrectly bonded (e.g., upside down) bonds. Orientational uniqueness is necessary to ensure physically symmetric parts are bonded in the desired orientation to allow complex mating of interconnecting structures.
In the prior art, bonding forces would be sufficient to dominate the mixing forces even if incorrectly aligned. In other words, certain incorrect orientations may result in a local minimization of energy, and the mixing energy is insufficient to move the part into the desired orientation to achieve the global energy minimum. Furthermore, for bonding approaches that have geometric symmetry, energy minimization may occur even in more than one physical orientation.
The gravity driven self-assembly process has attempted to overcome this issue. This is done by giving a specific shape to the receptor site hole. However, these approaches require parts with large scale asymmetrical physical geometries, adding cost and complexity to the batch manufacture process. In the capillary driven self-assembly method, asymmetric bonding interfaces have been implemented, but the alignment precision decreased and the process yields dropped by nearly 70%. The drop in yield is attributed to local energy minima creating misfits and the reduced precision was attributed to a less sharp dip in the energy curve.
The self-assembly method must also lend itself to bonding selectivity, i.e., that the desired part bond with its intended mating part, but not with a third unintended part. Each of the gravity, electrostatic, electric field trap, and capillary methods can be adapted to provide bonding selectivity based on geometric shape, electromagnetic properties, surface and hydrophilicity between the parts.
By way of example, for capillary driven assembly methods, a method has been developed for activating or deactivating certain receptor sites for the self-assembly of different components using sequential steps. Although allowing bonding selectivity, it requires substantial processing and precludes parallel assembly of heterogeneous mixtures. Furthermore, this sequential process lengthens the assembly process as a function of the number of different components.
For a gravity driven self-assembly, shape-matching techniques have been used to effect bonding selectivity for parallel self-assembly of a heterogeneous mixture of three different parts. It is comparable to the approach of a square block, which will not fit, into the proverbial circular hole. However, the number of mutually exclusive shapes may be limited and chip real estate may be wasted as a result of the need for size differentiation as one shape differentiator. Furthermore, the machining of arbitrarily shaped parts imposes additional processing complexity and cost. Therefore, the prior art provides no real solution for the bonding selectivity issue and the requirement in more complex applications for sequential selective bonding in a parallel process.
A self-assembly process should also enable inter-part bonding, namely that free floating parts bond to other free floating parts, rather than to fixed substrates. With respect to inter-part bonding, the electric field method becomes inapplicable with its requirement of the application of an electric charge to at least one of the bodies.
Inter-part bonding further requires the ability to bond in any arbitrary direction. Because of this, gravity-driven processes are inapplicable because gravity only acts in one direction. A collection of free floating parts would have no driving force to bond to one another unless oriented so that gravity is in the direction of at least one of the parts. As a result, short-range bonding forces must exist intrinsically between the parts, not from some externally applied source.
A secondary challenge for inter-part bonding is preventing agglomeration, where parts of a similar type inadvertently bond to each other rather than to the specified receptor site. As a result, capillary-driven assembly is inapplicable because agglomeration requires that there are no short range bonding forces between similar parts, while insuring there are bonding forces between dissimilar parts. If parts of a first type must have a wetted receptor, they will agglomerate to each other while the “dry” second type would not agglomerate.
For a capillary-driven self-assembly approach, it has been demonstrated that a two-step sequential self-assembly of encapsulated LED structures using shape- and solder-directed processes of free floating parts may overcome the inter-part bonding issue. The agglomeration problem is overcome by recessing the wetted receptor sites and cavities and limiting access to only the smaller parts that are intended to be bonded. Solder may also be used as the bonding liquid in a way that solid solder bumps are patterned at the wafer level and heated to melt to form wetted contacts on each of the individual components. When cooled, the solder also serves as the permanent electrical and mechanical contact. This process has been shown to have potential to be satisfactory, however, it suffers from the drawback that it requires sequential assembly steps and the limitation of a single electrical contact between the individual parts and the requirement for parts of dissimilar size to prevent agglomeration.
Accordingly, a method and structure for overcoming the shortcomings of the prior art and allowing self-assembly between two bodies while providing orientational uniqueness, bonding selectivity and inter-part bonding is desired.